Inscope plugin uses `saas-promotions` configmap created in backstage-prod namespace on appsre9 cluster, to render nodes on rollout diagram

With increasing number of complex saas pipelines, this configmap has grown in size and has crossed its capacity of 1MB. Currently a hack is put in place to omit 3 largest apps from rollout, in order to keep the rollouts functioning for other apps.

This design is not scalable, and this configmap needs to be deprecated, and instead the data should be grabbed straight from the s3 source file : i.e, publisher-data.json .

configmap query: https://gitlab.cee.redhat.com/service/app-interface/-/blob/master/resources/services/backstage/saas-promotions.configmap.yaml

configmap data source name: https://gitlab.cee.redhat.com/service/app-interface/-/blob/master/resources/services/backstage/saas-promotions.configmap.yaml#L6

AC:

• convert the jinja2 to a python http server that we will run as sidecar in the backstage plugin
  -it will periodically fetch data from s3 json file in the background
  and serve from memory (i.e. not making an s3 query during the request )
• this will serve as the data source to render rollout diagram

source: https://redhat-internal.slack.com/archives/C07FQ00BJV8/p1748595333003009?thread_ts=1748462991.403179&cid=C07FQ00BJV8

Risks:

• Concerns about correctness and completeness of data pushed by qontract to publisher-data.json file in s3. If it’s not accurate and complete, it should ne carded up in APPSRE.